Genes Involved in Differentiation, Stem Cell Renewal, and Tumorigenesis Are Modulated in Telomerase- Immortalized Human Urothelial Cells

Emma J. Chapman,1 Gavin Kelly,2 and Margaret A. Knowles1

1Cancer Research UK Clinical Centre, Leeds Institute of Molecular Medicine, St. James’s University Hospital, Leeds, United Kingdom and 2Bioinformatics and Biostatistics Service, Cancer Research UK, London Research Institute, Lincoln’s Inn Fields Laboratories, London, United Kingdom

Abstract non–telomere effects of telomerase and provides further The expression of hTERT, the catalytic subunit of rationale for the use of telomerase inhibitors in UC. telomerase, immortalizes normal human urothelial cells (Mol Cancer Res 2008;6(7):1154–68) (NHUC). Expression of a modified hTERT, without the ability to act in telomere maintenance, did not Introduction immortalize NHUC, confirming that effects at telomeres The primary and well-documented role of telomerase is as are required for urothelial immortalization. Previous a reverse transcriptase that acts in the maintenance of studies indicate that inhibition of telomerase has an telomere length and structure. Up-regulation of telomerase immediate effect on urothelial carcinoma (UC) cell line expression occurs in the majority of urothelial carcinoma viability, before sufficient divisions to account for (UC) irrespective or stage or grade (1), suggesting that this telomere attrition, implicating non–telomere effects may be an early event in tumorigenesis. Normal human of telomerase in UC. We analyzed the effects of urothelial cells (NHUC) are immortalized by expression of telomerase on gene expression in isogenic mortal and hTERT, the catalytic subunit of telomerase. In contrast to hTERT-transduced NHUC. hTERT expression led to requirements for immortalization in other epithelial cell types consistent alterations in the expression of genes and despite the common loss of expression of p16 in UC, predicted to be of phenotypic significance in inactivation of the CDKN2A locus (encoding p16 and tumorigenesis. A subset of expression changes were p14ARF) was not observed (2). detected soon after transduction with hTERT and Non–telomere effects of hTERT expression have been persisted with continued culture. These genes (NME5, described in other cell types, some of which may be relevant to PSCA, TSPYL5, LY75, IGFBP2, IGF2, CEACAM6, XG, tumorigenesis in vivo (3, 4). Inhibition of telomerase as a NOX5, KAL1 HPGD , and ) include eight previously therapeutic strategy is generally based on the assumption that identified as polycomb group targets. TERT-NHUC lack of telomerase activity will result in continued cell division showed overexpression of the polycomb repressor and telomere attrition, which will eventually lead to replicative complex (PRC1 and PRC4) components, BMI1 and senescence or apoptosis (5). However, inhibition of telomerase SIRT1, and down-regulation of multiple PRC targets and has an immediate effect on UC cell line viability, before genes associated with differentiation. TERT-NHUC at 100 sufficient divisions to account for telomere attrition (6). This population doublings, but not soon after transduction, strongly implicates non–telomere effects of hTERT in bladder showed increased saturation density and an attenuated tumorigenesis and suggests that telomerase inhibition may be of differentiation response, indicating that these are not rapid therapeutic benefit. Thus, identification of the genes and acute effects of telomerase expression. Some of the pathways involved in the non–telomere effects of telomerase in changes in gene expression identified may contribute bladder and other cancers may highlight novel therapeutic or to tumorigenesis. Expression of NME5 and NDN was diagnostic targets. There is also data that links the expression of down-regulated in UC cell lines and tumors. Our data telomerase with the inhibition of cellular differentiation (7, 8). supports the concept of both telomere-based and This may be a non–telomere event and is an example of how telomerase expression could contribute to tumorigenesis by mechanisms discrete from its classic actions in telomere Received 11/19/07; revised 4/18/08; accepted 4/21/08. maintenance. Grant support: In part by Cancer Research UK (C6228/A5433). hTERT-immortalized NHUC (TERT-NHUC) are generally The costs of publication of this article were defrayed in part by the payment of diploid and have no chromosomal alterations (detectable by page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. array CGH or karyotyping; ref. 2). However, changes in gene Note: Supplementary data for this article are available at Molecular Cancer expression after telomerase expression have not been Research Online (http://mcr.aacrjournals.org/). investigated. Microarray analysis of gene expression in Requests for reprints: Margaret A. Knowles, Cancer Research UK Clinical Centre, St. James’s University Hospital, Beckett Street, Leeds LS97TF, United isogenic mortal NHUC and their hTERT-immortalized coun- Kingdom. Phone: 44-11320-64913; Fax: 44-11324-29886. E-mail: m.a.knowles@ terparts was done to examine the hypothesis that expression of leeds.ac.uk Copyright D 2008 American Association for Cancer Research. telomerase contributes to tumorigenesis in ways that are doi:10.1158/1541-7786.MCR-07-2168 additional to its effect on telomere length and structure. As

TERT-NHUC had no detectable genetic alterations, we aimed it is possible that more subtle effects on telomere structure such to identify changes in gene expression that may have occurred as at the 3¶ overhang are required for immortalization (14). To via transcriptional mechanisms. Additionally, as TERT-NHUC determine whether immortalization of NHUC was due to effects provide the basis for an in vitro model of urothelial on telomere maintenance, cells were transduced to express transformation, it is important to determine whether alterations hTERT-HA. hTERT-HA has a carboxyl-terminal hemagglutinin in gene expression are present before other genes are (HA) tag and induces soluble telomerase activity but cannot act experimentally manipulated. in telomere maintenance, probably due to an inability to interact Several previous experiments have examined the effect of with the additional proteins required (15). Despite the induction hTERT on gene expression (9-12). There is little concordance of telomerase activity (7.2-fold compared with empty vector– between genes identified in these studies, which perhaps transduced cells), the expression of hTERT-HA did not lead to a reflects the cell type–specific pathways involved in immor- significant extension in the replicative life span of NHUC, talization. This is mirrored by the different combinations of confirming that telomere-dependent effects are required for the alterations seen in tumors arising within a particular tissue. immortalization of NHUC (Fig. 1). No previous study has examined changes in gene expression after telomerase expression in matched pairs of mortal and Telomerase Activity immortal epithelial cells from multiple donors. By repeating A low level of telomerase activity was detected in mortal the experiment in three biological replicates (derived from NHUC strains. Telomerase activity in each TERT-NHUC line three cell donors) and looking for changes in expression was quantified relative to that in the isogenic NHUC cell strain. consistent to multiple donors relative to their isogenic The ratios were 17, 9, and 6 for TERT-NHUC N, TERT-NHUC controls, inter-cell line differences should be minimal. This B, and TERT-NHUC A, respectively (Fig. 2A). experimental design should increase the power to detect genes whose expression is consistently altered after expres- Expression of hTERT Leads to Consistent and Stable sion of telomerase. We propose that the effects of telomerase Changes in Gene Expression in vivo are likely to be a combination and potentially Changes in gene expression of 2.0-fold or greater were synergistic effect of the classic actions in maintenance of identified following the comparison of hTERT-transduced and telomere length and structure coupled to its currently isogenic mortal strains of NHUC from three donors. This uncharacterized non–telomere effects. For this reason, we analysis was done soon after transduction with hTERT when chose to investigate the putative non–telomere effects of cells were still within their normal mortal life span [<18 telomerase in the biologically relevant context of fully population doublings (PD)] and also when cells were deemed functional telomerase. immortal and had undergone f100 and 250 PDs. The expression of hTERT led to a statistically significant down- Results regulation of 104 probe sets, early after transduction, in at Expression of a Modified hTERT without the Ability to least two of the three donors (Table 1). These comprised 87 Elongate Telomeres Does Not Immortalize NHUC genes and 17 unknown transcripts or open reading frames It has been unclear how expression of hTERT immortalizes (orf) or hypothetical genes. In cases in which a change in NHUC, as profound shortening of telomere length is not expression in cells from one of the three donors was not observed in NHUC at replicative senescence (2, 13). However, statistically significant, this was often due to a higher

FIGURE1. Transduction with TERT-HA does not immortalize NHUC, indicating that telomere-dependent effects are required for immortalization. *, wild-type hTERT; E, hTERT-HA; ., empty vector-transduced cells. Data is derived from cells seeded in triplicate wells and is repre- sentative of that obtained using cells from two independent donors.

FIGURE2. Characterization of TERT-NHUC. A. Quantification of telomerase activity in TERT-NHUC. Columns, mean ratio of telomerase activity relative to the empty vector – transduced isogenic NHUC. Data is derived from duplicate analyses. B. Relative quantification (RQ) of expression of BMI1 (gray columns), SIRT1 (black columns), and NDN (unshaded columns) in TERT-NHUC soon after transduction with hTERT. Expression is calculated relative to SDHA and normalized to a reference sample of pooled NHUC. C. The proposed role of NDN in controlling the expression of a subset of genes altered in TERT-NHUC. All interactions are known effects described in the literature. TERT-NHUC attenuated NDN expression. Down-regulation of NDN-dependent inhibition of E2F-1 could lead to an increase in the E2F-1 target gene BMI1. Up-regulation of BMI1 was detected in TERT-NHUC with down-regulated NDN. BMI1 forms part of the PRC1 which modulates the expression of a number of PCG targets. Multiple PCG targets were down-regulated in TERT-NHUC. D. Confirmation at 100 PD of down-regulated expression of NDN (unshaded columns) and up-regulated expression of BMI1 (black columns) in all three TERT- NHUC lines compared with the isogenic NHUC. Expression of NDN was undetectable after 40 cycles in TERT-NHUC B; this sample is assigned an arbitrary value of log10 (RQ) = À3.

variation in signal for that probe between the triplicate arrays component BMI1 was detected in cells from donor N. This for that donor or a fold change less than the stringent cutoff was confirmed by quantitative real-time reverse transcriptase of 2.0-fold. Examination of raw data often showed that the PCR (QRT-PCR; Fig. 2B). trend in gene expression was followed. Some genes modulated soon after the expression of Twenty-two of these 104 genes (indicated by asterisks in telomerase may be transient changes or those involved in a Table 1) have previously been identified as possible polycomb stress response following retroviral transduction. Therefore, gene (PCG) targets by Bracken et al. (16). Bracken et al. did a TERT-NHUC that had undergone f100 PD (the point at which genome-wide identification of human promoters bound by they were deemed immortal) were also examined for changes PCG and identified a list of >1,000 potential PCG targets. in gene expression. This second analysis (Table 2) showed that PCG proteins form multiprotein complexes called polycomb- 11 of the genes that were identified as acute alterations still repressive complexes (PRC) that play a key role in the showed consistent alteration. These genes (NME5, PSCA, regulation of transcription during development and differen- TSPYL5, LY75, IGFBP2, IGF2, CEACAM6, XG, NOX5, KAL1, tiation. SIRT1, a PRC4 component, was up-regulated in and HPGD) are therefore considered a ‘‘telomerase signature’’ cells from donors A and B and up-regulation of PRC1 of expression in NHUC. Of these 11 genes, 8 have been

described as PRC targets. This is a statistically significant The genes for which expression was no longer significantly (P = 0.0007, Fischer’s exact test) overrepresentation of PCG altered were CEACAM6, DNAJC15, GALNTL4, HOXC4, targets in the telomerase signature genes compared with the C1orf115, EXOSC6, TPST1, COL12A1, and NOS1.Ofthe11 other genes altered soon after telomerase expression. Modulat- ‘‘telomerase signature genes’’ all, with the exception of ed expression of these genes (with the exception of CEACAM6, were still significantly altered at this final analysis CEACAM6) is stable because these alterations persisted to time point. the final time point when cells had undergone f250 to 300 PD (Supplementary Table S1). Conditioned Medium from TERT-NHUC Does Not Affect At 100 PD, there was more consistency in the genes the Proliferation of NHUC modulated in TERT-NHUC. Twenty-two genes or transcripts, Unlike the effect reported in human mammary epithelial including hTERT, were modulated in all three TERT-NHUC cells (3), expression of telomerase in NHUC did not lead to the lines, and apart from hTERT, all were down-regulated (Table alteration of expression of growth factors or receptors such as 2). This suggests that a general mechanism of gene repression fibroblast growth factor and epidermal growth factor receptors. had been activated. The gene that showed most consistent In accordance with this, conditioned medium from TERT- (all three lines) and profound down-regulation in TERT- NHUC had no effect on the proliferation of unmodified NHUC NHUC at 100 PD was necdin (NDN; ref. 17). A significant (data not shown). down-regulation of NDN was not detected soon after the transduction with hTERT by array analysis. However, QRT- Induction of Differentiation in TERT-NHUC PCR showed some down-regulation of NDN in early passage As several genes that were consistently down-regulated in TERT-NHUC N (but not in TERT-NHUC B or TERT-NHUC TERT-NHUC at 100 PD were associated with differentiation A; Fig. 2B). We hypothesize that down-regulation of NDN, in (genes identified in italics in Tables 1 and 2), it was of interest turn, leads to the down-regulation of a significant subset of to determine whether these cells retained a normal differenti- genes in NHUC with long-term expression of hTERT (Fig. ation response. Peroxisome proliferator–activated receptor-g 2C). As it is known to bind to E2Fs and repress E2F- (PPARg) signaling is involved in urothelial differentiation. dependent transcription (18), down-regulation of NDN is Previously, treatment of NHUC with the PPARg agonist predicted to lead to the increased expression of BMI1, a Troglitazone, together with the inhibition of autocrine epider- known E2F target gene (19). Up-regulation of BMI1 was mal growth factor receptor signaling by the small molecule identified on array analyses of TERT-NHUC at 100 PD inhibitor, PD153035, has been shown to induce the expression and was confirmed by QRT-PCR in cells from all three of urothelial differentiation–associated markers, uroplakin II donors (Fig. 2D). Of note, TERT-NHUC N, the only cell line (UPK2) and cytokeratin 20 (CK20; refs. 20-22). Treatment with to show early NDN down-regulation soon after transduction PD153035 alone has no effect on UPK2 expression (20). We with hTERT, also showed BMI1 up-regulation. BMI1 forms assessed the induction of CK20 expression by immunofluores- part of the PRC1 and many of the genes (highlighted by cence microscopy in NHUC and found that as induction was asterisks in Tables 1 and 2) are PCG targets as identified by restricted to a minority of cells in the culture (data not shown) Bracken et al. (16). and there was interdonor variability, this assay was not As the expression of telomerase is an early event in sufficiently quantitative for the assessment of differentiation tumorigenesis and its expression persists from premalignancy in TERT-NHUC. Therefore, the ability to differentiate was to tumor development, we were interested in the changes in assessed by QRT-PCR for UPK2. Treatment of all three TERT- gene expression that persisted in hTERT-immortalized cells as NHUC lines at the 100 PD time point with Troglitazone and these may confer a selective advantage and might be PD153035 resulted in an increase in UPK2 expression. In the important in tumorigenesis. Examination of the gene expres- case of TERT-NHUC N and TERT-NHUC B, isogenic NHUC sion signature of TERT-NHUC at both <18 PD and 100 PD were available and we observed that the magnitude of UPK2 time points using Ingenuity Pathway Analysis software induction was less than that seen in NHUC, indicating that showed that cancer was the disease or disorder most the response is attenuated following immortalization by closely associated with many of the alterations in gene telomerase (Fig. 4A). Induction of UPK2 was then measured expression (Fig. 3). This confirmed our observation in in TERT-NHUC soon after the expression of telomerase in which altered expression of several of these genes has comparison to those that had undergone f100 PD (Fig. 4B). previously been associated with bladder or other cancers, In these cells, the induction of UPK2 was greater than in e.g., CXADR and tumor suppressor candidates such as NDN those analyzed at 100 PD and was similar to that previously and NME5. Analysis of gene ontology groups identified the observed in isogenic NHUC. This suggests that attenuation overrepresentation of 15 genes involved in 31 often over- in the ability to differentiate is not an acute effect of telomerase lapping gene ontology categories in these cells (Supplemen- but rather is related to changes in gene expression that were tary Table S2). detected after continued proliferation. Morphologic changes Expression array analysis was done on TERT-NHUC that after treatment with Troglitazone and PD153035 have been had undergone an additional period of proliferation of at least described in NHUC (22). Neither early nor later passage TERT- 150 PD (Supplementary Table S1). This third analysis NHUC showed the characteristic ‘‘grosetteh’’ morphology confirmed that all but nine of the genes altered at the 100 PD observed in NHUC (Fig. 4C), indicating that even soon after time point still showed altered expression (>2.0-fold in cells telomerase expression, differentiation may be attenuated to from at least two donors) after this prolonged culture period. some extent.

Table 1. Genes Whose Expression Was Modulated in NHUC Soon (<18 PD) after Expression of hTERT

Probe set TERT-NHUC Symbol Description

ABN

Down-regulated genes 204351_at À46.71 À43.58 S100P S100 calcium-binding protein P 39248_at À16.12 À9.64 AQP3 Aquaporin 3 (Gill blood group) 231259_s_at À13.00 À5.69 CCND2* Cyclin D2 205668_at À4.27 À12.09 LY75* Lymphocyte antigen 75 202718_at À5.19 À10.44 IGFBP2* Insulin-like growth factor binding protein 2, 36 kDa 204818_at À7.23 À8.17 HSD17B2 Hydroxysteroid (17-h) dehydrogenase 2 202409_at À10.73 À9.22 IGF2* Insulin-like growth factor 2 (somatomedin A) 211657_at À8.10 À4.45 À6.86 CEACAM6* Carcinoembryonic antigen-related cell adhesion molecule 6 231169_at À4.50 À6.30 À7.73 202295_s_at À4.35 À5.58 CTSH Cathepsin H 1554062_at À11.71 À2.82 XG* Xg blood group 229352_at À10.11 À4.07 NOX5* NADPH oxidase, EF-hand calcium binding domain 5 227554_at À3.33 À5.96 MRNA; cDNA DKFZp686I18116 206197_at À3.76 À5.57 À3.74 NME5 Nonmetastatic cells 5, protein expressed in prostate stem cell antigen 205319_at À4.22 À4.29 PSCA 242277_at À7.32 4.44 PHACTR2 Phosphatase and actin regulator 2 230921_s_at À3.19 À4.26 À3.60 211548_s_at Hydroxyprostaglandin dehydrogenase 15-(NAD) 203914_x_at 203913_s_at À3.97 À5.29 HPGD* 222240_s_at À3.04 À3.17 ISYNA1 myo-inositol 1-phosphate synthase A1 213122_at À3.27 À2.83 TSPYL5 TSPY-like 5 220494_s_at À3.00 À3.04 242844_at À2.19 À3.12 À3.66 PGGT1B Protein geranylgeranyltransferase type I, h subunit 240277_at À3.21 À2.72 SLC30A7 Solute carrier family 30 (zinc transporter), member 7 225730_s_at À3.25 À2.53 THUMPD3 THUMP domain containing 3 229128_s_at À5.14 À3.39 ANP32E Acidic (leucine-rich) nuclear phosphoprotein 32 family, member E 228443_s_at À2.89 À2.57 SETD8 SET domain containing (lysine methyltransferase) 8 235863_at À2.46 À2.94 JSRP1 Junctional sarcoplasmic reticulum protein 1 241669_x_at À2.50 À2.80 PRKD2 Protein kinase D2 207306_at À2.23 À3.04 TCF15* Transcription factor 15 (basic helix-loop-helix) 213872_at À4.12 À3.78 C6orf62 Chromosome 6 open reading frame 62 227338_at À2.75 À2.52 LOC440983 Hypothetical gene supported by BC066916 238326_at À3.66 À4.22 LOC440836 Similar to MGC52679 protein 240757_at À3.79 À4.05 CLASP1 Cytoplasmic linker associated protein 1 214182_at À2.58 À2.63 ARF6 ADP-ribosylation factor 6 213747_at À4.58 À3.19 AZIN1 Antizyme inhibitor 1 235186_at À2.68 À2.47 LOC644634 Hypothetical LOC644634 222614_at À4.49 À3.22 RWDD2B RWD domain containing 2B 238267_s_at À2.45 À2.68 213710_s_at À4.59 À3.07 211948_x_at 214055_x_at À2.35 À2.51 À2.66 BAT2D1 BAT2 domain containing 1 208994_s_at À2.51 À2.47 PPIG Peptidylprolyl isomerase G (cyclophilin G) 227473_at À3.78 À3.64 231268_at À2.27 À2.61 LOC645895 Hypothetical LOC645895 201120_s_at À2.43 À2.41 PGRMC1 Progesterone receptor membrane component 1 241671_x_at À2.43 À2.41 FLJ22536 Hypothetical locus LOC401237 240686_x_at À2.33 À2.50 TFRC Transferrin receptor (p90, CD71) 215454_x_at À2.32 À2.49 SFTPC* Surfactant, pulmonary-associated protein C 229676_at À3.90 À3.29 PAPD1 PAP associated domain containing 1 229713_at À3.79 À3.37 217696_at À2.15 À2.62 FUT7 Fucosyltransferase 7 (a (1,3) fucosyltransferase) 206042_x_at À2.18 À2.55 SNRPN upstream reading frame 243297_at À2.27 À2.39 VPS13D Vacuolar protein sorting 13 homologue D (S. cerevisiae) 229747_x_at À2.29 À2.38 MGC40489 Hypothetical protein MGC40489 241618_at À2.32 À2.27 PACS1 Phosphofurin acidic cluster sorting protein 1 218094_s_at À2.41 À2.17 DBNDD2 Dysbindin (dystrobrevin binding protein 1) domain containing 2 217040_x_at À2.00 À2.57 SOX15 SRY (sex determining region Y)-box 15 210733_at À3.05 À3.69 TRAM1 Translocation associated membrane protein 1 214630_at À2.34 À2.07 CYP11B2* Cytochrome P450, family 11, subfamily B, polypeptide 2 205212_s_at À2.22 À2.15 CENTB1* Centaurin, h1 1555226_s_at À2.05 À2.23 C1orf43 Chromosome 1 open reading frame 43 215778_x_at À2.15 À2.11 HAB1 B1 for mucin 227897_at À3.87 À2.53 RAP2B RAP2B, member of RAS oncogene family 238051_x_at À2.01 À2.22 PWWP2 PWWP domain containing 2 217410_at À2.05 À2.17 AGRN Agrin 227781_x_at À2.01 À2.19 FAM57B Family with sequence similarity 57, member B Up-regulated genes 202237_at 37.34 11.82 NNMT* Nicotinamide N-methyltransferase 1555271_a_at 12.59 8.68 12.91 TERT Telomerase reverse transcriptase

Continued on the following page

Table 1. Genes Whose Expression Was Modulated in NHUC Soon (<18 PD) after Expression of hTERT (Cont’d)

Probe set TERT-NHUC Symbol Description

AB N

235057_at 5.71 5.67 8.35 ITCH* Itchy homologue E3 ubiquitin protein ligase 206230_at 4.37 7.46 LHX1* LIM homeobox 1 244075_at 3.88 7.64 HSD17B7P2 Hydroxysteroid (17-h) dehydrogenase 7 pseudogene 2 218878_s_at 3.05 7.53 SIRT1 Sirtuin 232174_at 7.02 6.93 EXT1 Exostoses (multiple) 1 205932_s_at 3.14 5.41 MSX1* msh homeobox 1 229574_at 6.59 4.90 TRA2A Transformer-2a 215505_s_at 3.06 5.78 2.62 STRN3 Striatin, calmodulin binding protein 3 238735_at 3.80 3.54 TCF12 Transcription factor 12 204255_s_at 7.81 2.58 VDR* Vitamin D (1,25-dihydroxyvitamin D3) receptor 204837_at 3.39 3.25 MTMR9* Myotubularin-related protein 9 209811_at 34449_at 3.50 3.11 CASP2 Caspase 2 1553111_a_at 2.49 3.44 KBTBD6 Kelch repeat and BTB (POZ) domain containing 6 219235_s_at 3.33 2.49 PHACTR4 Phosphatase and actin regulator 4 240145_at 2.87 2.92 DGKH* Diacylglycerol kinase, q 207078_at 2.98 2.73 MED6 Mediator complex subunit 6 204669_s_at 2.01 3.42 RNF24 Ring finger protein 24 203243_s_at 2.56 2.83 PDLIM5 PDZ and LIM domain 5 204730_at 3.39 4.66 RIMS3 Regulating synaptic membrane exocytosis 3 205206_at 1.93 3.15 KAL1* Kallmann syndrome 1 sequence 225484_at 2.51 2.55 TSGA14 Testis-specific, 14 236816_at 2.41 2.60 C12orf30 Chromosome 12 open reading frame 30 223679_at 2.59 2.37 CTNNB1* Catenin (cadherin-associated protein), h1, 88 kDa 239709_at 2.51 2.45 RP11-78J21.1 Heterogeneous nuclear ribonucleoprotein A1-like 235432_at 2.78 2.16 NPHP3 Nephronophthisis 3 (adolescent) 230779_at 4.38 3.03 TNRC6B Trinucleotide repeat containing 6B 224595_at 4.29 3.08 SLC44A1 Solute carrier family 44, member 1 238738_at 4.33 2.80 PSMD7 Proteasome 26S subunit, non-ATPase, 7 204115_at 2.06 2.66 GNG11 Guanine nucleotide binding protein g11 222557_at 3.17 3.77 STMN3 Stathmin-like 3 202723_s_at 3.19 3.66 FOXO1* Forkhead box O1 203355_s_at 2.75 4.04 PSD3 Pleckstrin and Sec7 domain containing 3 208200_at 2.29 2.21 2.24 IL1A Interleukin 1a 225662_at 4.09 2.61 ZAK Sterile a motif and leucine zipper containing kinase AZK 214806_at 3.57 2.89 BICD1 Bicaudal D homologue 1 (Drosophila) 201040_at 2.07 2.04 GNAI2 Guanine nucleotide binding protein, a-inhibiting activity polypeptide 2 225961_at 2.84 3.32 KLHDC5 Kelch domain-containing 5

TERT-NHUCShow Increased CultureSaturation Density regulated in the majority of UC cell lines compared with pooled TERT-NHUC, soon after transduction, had an average NHUC. NDN was down-regulated in 26 of 28 (92.9%; saturation density (the confluent cell density at which cell Fig. 6A), NME5 in 27 of 28 (96.4%; Fig. 6B), and ADFP in proliferation is contact-inhibited) of 1.18 Â 105 cells/cm2, 16 of 28 (57.1%; Fig. 6C). Down-regulation of NDN protein which is similar to that of the isogenic NHUC and the expression in cell lines was confirmed by Western blotting (data previously published value for NHUC of 1 Â 105/cm2 (ref. 23; notshown).ExpressionofNDNandNME5wasthen Fig. 5A). However, TERT-NHUC at 100 PD had a higher investigated in a panel of primary UC. NDN was down- saturation density compared with their paired isogenic NHUC regulated in 35 of 58 (60%; Fig. 6D) and NME5 in 10 of 47 and on average, reached contact inhibition at 1.7 Â 105 cells/cm2 (21%; Fig. 6E), demonstrating that changes in the expression of compared with 0.91 Â 105 cells/cm2 for NHUC (Fig. 5B). genes that are modulated in TERT-NHUC also occur in bladder Genes involved in cell-cell signaling, such as ICAM2 were cancer in vivo. down-regulated in these cells, which may have contributed to this phenotype. Discussion Genes Down-Regulated in TERT-NHUCAre Also Down- We have shown previously that expression of hTERT Regulated in UCCellLines and Tumors immortalizes NHUC in vitro with no detectable chromosomal NDN, NME5, and ADFP were selected for further analyses alterations. It was not clear whether immortalization was due to on the basis of profound and consistent down-regulation in cells the telomere-dependent effects of telomerase, as a low level of at 100 PD. As discussed later, these genes have potential tumor endogenous telomerase activity is detected in cultured NHUC, suppressor functions and are implicated in mediating cellular and profound telomere length-shortening is not seen at differentiation. Expression was investigated in UC cell lines replicative senescence (13). However, the expression of a using TaqMan QRT-PCR. Expression of each gene was down- modified hTERT (hTERT-HA), that retains telomerase activity

Table 2. Genes Whose Expression Was Modulated in TERT-NHUC at 100 PD

Probe set TERT-NHUC Symbol Description

ABN

Down-regulated genes 209122_at À64.52 À35.12 ADFP Adipose differentiation-related protein 208596_s_at À49.41 À14.48 UGT1A3 UDP glucuronosyltransferase 1 family, polypeptide A3 202409_at À5.74 À45.69 À35.40 LOC492304 NA 209550_at À26.41 À23.44 À28.06 NDN Necdin homologue (mouse) 215125_s_at 206094_x_at À39.47 À12.36 UGT1A6* UDP glucuronosyltransferase 1 family, polypeptide A6 215440_s_at À26.23 À21.15 BEXL1 Brain expressed X-linked-like 1 214974_x_at 215101_s_at À16.09 À21.40 CXCL5* Chemokine (C-X-C motif) ligand 5 202157_s_at À7.06 À27.65 CUGBP2 CUG triplet repeat, RNA binding protein 2 221024_s_at À18.02 À16.26 SLC2A10 Solute carrier family 2 member 10 218694_at À17.65 À15.59 ARMCX1 Armadillo repeat-containing, X-linked 1 203917_at 239155_at À10.55 À32.8 À6.37 CXADR* Coxsackie virus and adenovirus receptor 205513_at À10.50 À22.27 TCN1 Transcobalamin I 207126_x_at À20.53 À11.30 UGT1A1 UDP glucuronosyltransferase 1 family, polypeptide A1 204532_x_at À20.89 À8.99 UGT1A9 UDP glucuronosyltransferase 1 family, polypeptide A9 202075_s_at À16.15 À11.20 PLTP Phospholipid transfer protein 204083_s_at À16.17 À10.34 TPM2 Tropomyosin 2 (h) 201348_at 214091_s_at À10.33 À16.33 GPX3 Glutathione peroxidase 3 (plasma) 204417_at À16.13 À9.38 GALC* Galactosylceramidase (Krabbe disease) 235940_at À14.27 À9.82 C9orf64 Chromosome 9 open reading frame 64 216623_x_at 215108_x_at À16.36 À4.12 TNRC9 Trinucleotide repeat containing 9 212859_x_at À6.81 À12.53 MT1E* Metallothionein 1E (functional) 205856_at 229151_at À15.67 À11.72 SLC14A1 Solute carrier family 14 (urea transporter), member 1 206714_at 1555416_a_at À9.87 À8.94 ALOX15B Arachidonate 15-lipoxygenase, second type 206197_at À11.88 À5.67 À9.94 NME5 Nonmetastatic cells 5, protein expressed in 238021_s_at À3.36 À14.61 LOC388279 NA 202888_s_at À14.98 À2.43 ANPEP Alanyl (membrane) aminopeptidase 205997_at À11.33 À5.57 ADAM28* ADAM metallopeptidase domain 28 202718_at À5.27 À11.19 IGFBP2* Insulin-like growth factor binding protein 2, 36kDa 211657_at À10.37 À5.33 CEACAM6* Carcinoembryonic antigen-related cell adhesion molecule 6 229352_at À9.85 À5.64 NOX5* NADPH oxidase, EF-hand calcium binding domain 5 205668_at À5.20 À12.10 À5.33 LY75* Lymphocyte antigen 75 224435_at À7.51 À6.12 C10orf58* Chromosome 10 open reading frame 58 1554062_at À2.86 À11.70 À4.61 XG* Xg blood group 202410_x_at À8.02 À4.62 IGF2* Insulin-like growth factor 2 (somatomedin A) 1552566_at À3.99 À8.62 C10orf87 Chromosome 10 open reading frame 87 203921_at À5.75 À6.86 CHST2 Carbohydrate sulfotransferase 2 229095_s_at À9.36 À6.71 À2.67 LIMS3 LIM and senescent cell antigen-like domains 3 221690_s_at À6.27 À5.69 NALP2 NACHT, leucine rich repeat and PYD containing 2 1569110_x_at À5.38 À6.56 PDCD6 Programmed cell death 6 221950_at À5.58 À5.42 À6.52 EMX2* Empty spiracles homologue 2 (Drosophila) 228843_at À7.64 À3.29 À6.03 NA 213122_at À7.53 À3.04 À5.68 TSPYL5 TSPY-like 5 228080_at À8.18 À4.00 À3.29 LOC143903 NA 204984_at À3.63 À6.61 GPC4 Glypican 4 205501_at À7.38 À2.79 NA 1559827_at À5.58 À4.50 LOC401074 NA 239082_at À5.23 À4.73 NA 211732_x_at 204112_s_at À4.04 À5.51 HNMT Histamine N-methyltransferase 203274_at À6.70 À2.40 F8A1 Coagulation factor VIII – associated (intronic transcript) 1 203423_at À5.32 À3.67 RBP1 Retinol binding protein 1, cellular 223824_at À5.48 À3.95 À3.94 C10orf59 Chromosome 10 open reading frame 59 210664_s_at À4.04 À6.13 À3.03 TFPI Tissue factor pathway inhibitor 223960_s_at À3.40 À5.46 C16orf5 Chromosome 16 open reading frame 5 205493_s_at À4.13 À4.65 DPYSL4 Dihydropyrimidinase-like 4 222592_s_at 218322_s_at À5.91 À2.79 ACSL5 Acyl-CoA synthetase long-chain family member 5 1555564_a_at À5.07 À2.85 IF* I factor (complement) 205319_at À5.20 À3.71 À2.49 PSCA Prostate stem cell antigen 213620_s_at À4.95 À3.31 À3.11 ICAM2 Intercellular adhesion molecule 2 1554079_at À2.30 À5.04 GALNTL4 UDP-N-acetyl-a-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase-like 4 214077_x_at À3.70 À3.20 MEIS4 Meis1, myeloid ecotropic viral integration site 1 homologue 4 231728_at 231729_s_at À3.81 À2.99 CAPS Calcyphosine 223832_s_at À2.72 À2.78 À4.78 CAPNS2 Calpain, small subunit 2 203192_at À3.79 À3.02 ABCB6 ATP-binding cassette, sub-family B (MDR/TAP), member 6 227554_at À3.68 À3.10 LOC402560 NA 218435_at 227808_at À3.64 À2.95 DNAJC15 DnaJ (Hsp40) homologue, subfamily C, member 15 205073_at À3.41 À3.11 CYP2J2* Cytochrome P450, family 2, subfamily J, polypeptide 2 203404_at À3.79 À2.66 ARMCX2* Armadillo repeat-containing, X-linked 2 211548_s_at À3.84 À2.45 HPGD* Hydroxyprostaglandin dehydrogenase 15-(NAD) 203913_s_at 203914_x_at

Continued on the following page

Table 2. Genes Whose Expression Was Modulated in TERT-NHUC at 100 PD (Cont’d)

Probe set TERT-NHUC Symbol Description

ABN

227006_at À2.48 À2.73 À4.22 PPP1R14A Protein phosphatase 1, regulatory (inhibitor) subunit 14A 220911_s_at À3.25 À3.06 À2.75 KIAA1305 KIAA1305 227803_at À2.32 À3.60 ENPP5* Ectonucleotide pyrophosphatase/phosphodiesterase 5 242856_at 4.52 À7.70 À5.62 NA 203799_at À3.32 À2.44 CD302 CD302 antigen 218094_s_at À3.33 À2.29 C20orf35* chromosome 20 open reading frame 35 227892_at À2.45 À3.16 NA 228721_at À3.20 À2.65 À2.27 LOC339903 NA 206194_at À2.62 À2.69 HOXC4 Homeo box C4 220027_s_at À2.12 À2.76 À2.61 RASIP1* Ras interacting protein 1 225841_at À2.15 À2.17 C1orf59 Chromosome 1 open reading frame 59 205827_at À2.10 À1.93 CCK Cholecystokinin Up-regulated genes 205952_at 40.00 26.67 KCNK3* Potassium channel, subfamily K, member 3 1555271_a_at 8.14 24.16 9.30 TERT Telomerase reverse transcriptase 238846_at 15.67 10.07 TNFRSF11A* Tumor necrosis factor receptor superfamily, member 11a 227307_at 5.18 13.49 TSPAN18 Tetraspanin 18 202265_at 8.02 9.50 BMI1 Polycombgroup ring finger 4 (BMI1) 203895_at 12.93 2.41 PLCB4 Phospholipase C, h4 227641_at 5.31 8.84 FBXL16* F-box and leucine-rich repeat protein 16 202555_s_at 224823_at 5.47 7.49 MYLK Myosin, light polypeptide kinase 239132_at 5.45 4.43 NOS1 Nitric oxide synthase 1 (neuronal) 209493_at 4.02 5.42 PDZK3 PDZ domain containing 3 227123_at 4.40 3.93 RAB3B RAB3B, member RAS oncogene family 231766_s_at 3.20 4.43 COL12A1 Collagen, type XII, a1 205206_at 2.81 4.60 KAL1* Kallmann syndrome 1 sequence 202957_at 4.21 3.19 HCLS1 Hematopoietic cell-specific Lyn substrate 1 204140_at 3.65 3.75 TPST1 Tyrosylprotein sulfotransferase 1 231916_at 3.62 3.73 EXOSC6 Exosome component 6 218546_at 2.23 2.00 C1orf115 Chromosome 1 open reading frame 115

NOTE: Genes are ranked in order of average fold change. Genes previously identified as polycombtargets are highlighted with asterisks, and telomera se signature genes are in boldface. Genes with associations with differentiation are shown in italics. but is deficient in the ability to elongate telomeres, failed to reportedly PCG targets. As it is estimated that between 1% and confer any significant extension in replicative life span, 5% of all genes are PCG targets (25), there is a significant bias confirming that the actions of telomerase in telomere towards PCG target genes in this expression profile. TERT- maintenance are required for the immortalization of NHUC. NHUC, soon after transduction with hTERT, had overexpres- Recently, Choi et al. have described the effects of a reverse sion of SIRT1, a NAD+-dependent deacetylase. SIRT1 with transcriptase–defective hTERT in transcriptional regulation of EED2 forms part of PRC4 (26) and is involved in epigenetic multiple genes converging on developmental pathways in skin silencing by PCG proteins (27) and aberrant methylation of progenitor cells (24). This and data presented here supports the tumor suppressor proteins (28). SIRT1 promotes transcriptional concept of non–telomere effects of hTERT on gene expression. repression by deacetylating specific histone proteins, recruiting Identification of changes in gene expression that occur soon histone H1band modulating the activity of SUV39H1 (29), the after transduction with hTERT and persist with continued enzyme responsible for the accumulation of trimethylated culture of TERT-NHUC identifies genes which can be histone H3 (H3K9me) in a region of chromatin. Thus, SIRT1 is considered a telomerase signature of gene expression. These a good candidate for causing acute telomerase-associated are genes that could potentially be directly modulated by the modulation of gene expression, although the mechanism by expression of telomerase. Our data, for the first time, identifies which telomerase expression may result in up-regulation of the involvement of polycombgene pathways. As TERT-NHUC SIRT1 is unknown. had no detectable chromosomal alterations and these changes in We believe that examination of gene expression at 100 PD expression occur in a timescale which makes spontaneous (in an immortalized but nontransformed cell population) may mutation unlikely, we suggest that this telomerase signature is identify those changes in gene expression that confer a of epigenetic origin. The fact that not all those alterations phenotypic advantage and that may be relevant to tumorigen- identified soon after transduction with hTERT persist with esis in vivo. Changes in gene expression at this time point are continued culture supports the concept that these changes are stable as further microarray expression analysis after a due to transcriptional rather than permanent genetic changes. prolonged culture period found that nearly all changes in However, the possibility that a proportion of the changes in expression were still present. TERT-NHUC have no identifiable gene expression are due to currently unidentified mutations chromosomal alterations, and therefore, it is likely that these cannot be discounted. genes have been silenced by transcriptional or epigenetic Twenty-two of the 104 (21%) transcripts whose expression mechanisms. At this second time point, there was again a was modulated soon after the expression of telomerase are high number of modulated genes described as PCG targets. The

FIGURE3. Ingenuity Pathway Functional Analysis identified the diseases and disorders (A) and cellular functions (B) that were most significant to the gene expression signature set of TERT-NHUC. Vertical line on each chart, threshold level of P = 0.05. Data from cells soon after transduction with hTERT (<18 PD; black columns) and data from cells at 100 PD (gray columns).

most down-regulated gene in all three TERT-NHUC lines at growth suppressor in postmitotic neurons (34), is silenced in 100 PD was NDN. NDN is of interest as it is both a novel neuroblastoma (35), and has roles in differentiation (36, 37). candidate tumor suppressor gene and a potential modulator NDN is involved in the interaction of nerve growth factor of a subset of other changes in gene expression via its inter- with its receptor p75NTR (18, 38). p75NTR signaling is action with E2F1, and consequently, BMI1 and PRC1 target implicated in the control of epithelial cell growth and genes (Fig. 2C). We found that the expression of NDN was differentiation (39), and induction of apoptosis in bladder down-regulated in a high proportion of UC examined sug- cells (40). NDN binds to and represses the activity of SV40 gesting that this may indeed be relevant to tumor development large T (34, 38). It also interacts with p53 (41), antagonizes in vivo. E2F1-mediated transcription, inhibits apoptosis, and sup- NDN maps within an imprinted region on 15q11 presses colony formation of osteosarcoma cells (18, 34, 41, implicated in the pathogenesis of the neurodevelopmental 42). NDN also directly binds to specific DNA sequences and disorder Prader-Willi syndrome, where it is silenced by acts as a transcriptional repressor (43). deletion, maternal uniparental disomy or translocation. Several We hypothesize that down-regulation of NDN releases the observations suggest that NDN has a tumor suppressor role. inhibition of E2F1-dependent transcription of BMI1. BMI1 Although not currently recognized as a cancer-prone syn- forms part of PRC1, which is a chromatin-modifying complex drome, an increased risk of leukemia has been reported in involved in the control of gene expression and is implicated in Prader-Willi syndrome (30). There are also reports linking stem cell renewal (44) and in delaying senescence (45). Prader-Willi syndrome with solid tumors (31-33). NDN is a Twenty-five of the genes altered in TERT-NHUC at 100 PD

are putative polycombgene targets (16), as are 8 of the 11 genes defined as the telomerase signature that was induced and persisted throughout the period of study. Up-regulation of BMI1 and other polycombgenes occurs in a range of tumor types (46) and precancerous tissue (47), in which they are thought to promote tumorigenesis by transcriptional repression of tumor suppressor genes and possibly via effects on stem cell maintenance (48). BMI1 is also associated with a stem-like expression profile that predicts poor outcome and treatment failure in multiple tumor types including bladder cancer (49). BMI1 expression is required for immortalization of some cell types by telomerase, possibly due to its effects in silencing p16/p14ARF transcription. Up-regulation of BMI1 was not associated with the attenuation of p16 or p14ARF transcription in TERT-NHUC. However, it is possible that BMI1 expression contributed to the immortalization of NHUC by preventing significant up-regulation of p16 expression, which plays a role in the control of the NHUC replicative life span. Although not a widely acknowledged function of telomerase, other studies have shown a reciprocal relationship between telomerase activity and differentiation (7, 8, 50-52). Bracken et al. (16) reported that many putative polycombgene targets such as those identified here, have roles in differentiation and development. TERT-NHUC at 100 PD showed reduced expression of many genes associated with differentiation. In many cases, there is evidence that down-regulation of these differentiation-associated genes also occurs in cancer. For example, the expression of CXADR, which was consistently down-regulated in TERT-NHUC, is known to be significantly reduced in invasive compared with superficial bladder cancers (53). A dramatic down-regulation of UDP-glucuronosyltransferase family of detoxifying enzymes was observed in TERT- NHUC from donors N and B. In normal bladder, UGT staining seems to correlate with epithelial cell differentiation and is decreased in some UCs (54). Also, prostate stem cell antigen is widely expressed in normal urothelium and noninvasive urothelial tumors, and is down-regulated in undifferentiated bladder carcinomas, leading to its description as a potential molecular marker of dedifferentiation in urothelial cells (55). ALOX15B is also of interest as one of its products, 15-S- hydroxyecosatetraenoic acid, an endogenous ligand for PPARg, is known to be pivotal in urothelial cell differentiation (20). Its expression in mature squamous but not basal keratinocytes also hints at a role in cellular differentiation and it is down-regulated in various primary tumors and cell lines (56). Several other genes that were down-regulated after the expression of telomerase have not yet been shown to have roles in urothelial cell differentiation, but there is evidence that they FIGURE4. A. Expression of differentiation-associated uroplakin II may play this role in other cell types. NDN has roles in the was induced in TERT-NHUC (at 100 PD) after treatment with the differentiation of smooth muscle cells, adipocytes, and neurons PPARg agonist Troglitazone and the epidermal growth factor (36, 37, 57). NME5 is a homologue of nm23-H1, a tumor receptor inhibitor PD153035. The magnitude of the response was suppressor previously linked to bladder cancer (58). The less than that seen in isogenic NHUC. Data shows log10 relative quantification (RQ) relative to SDHA control gene and normalized to function of NME5, also known as nm23 H5, is not well- pooled NHUC cDNA. Expression data was derived from the average described, although it is implicated in the differentiation of of duplicate experiments. B. Comparison of induction of UPK2 in TERT-NHUC at early and late passage shows that attenuation of spermatozoa (59). We detected altered NME5 expression in differentiation is not an acute effect of telomerase expression. C. UC. To our knowledge, this is the first investigation of NME5 Troglitazone and PD153035 treatment of TERT-NHUC did not result in the characteristic ‘‘rosette’’ morphology seen in NHUC (arrow). expression in any tumor type. Expression of LY75 (gp200- Bars, 200 Am. MR6) has been linked to the differentiation of colorectal cell

FIGURE5. A. Maximum saturation density obtained in TERT-NHUC soon after transduction with hTERT (<18 PD) was less than that seen in TERT- NHUC that had undergone f100 PD. TERT-NHUC soon after transduction with hTERT had a saturation density similar to that of NHUC. Columns, range of values within triplicate donors. B. TERT-NHUC (at 100 PD) reached a higher density before contact inhibition compared with their matched NHUC. Points, mean of cells from three donors, values from triplicate donors: filled symbols, TERT-NHUC (at f100 PD); open circles, NHUC; bars, SD.

lines (60). ADFP is a transcriptional target of PPARg.Its bladder and other cancers is merited. Genes altered after acute expression is decreased in undifferentiated renal cell carcinoma, and prolonged expression of telomerase include both PRC and ADFP-positive tumors are associated with improved components and PCG target genes. These non–telomere survival (61). actions of hTERT may explain the predominance of activation As TERT-NHUC at 100 PD had a ‘‘de-differentiated’’ gene of telomerase in bladder and other epithelial cancers rather than expression profile, it was of interest to determine whether the alternative lengthening of telomeres’ pathway. TERT-NHUC such as hTERT-immortalized bronchial epithelial These considerable alterations in gene expression should not cells (62) retain the ability to differentiate. TERT-NHUC (and preclude the use of TERT-NHUC as experimental tools NHUC) responded to a PPARg agonist with the induction of in vitro but do argue for the consideration of these changes UPK2. However, the fold-change of induction of UPK2 in when using these in place of normal unmodified cells. We TERT-NHUC at 100 PD was less than that seen in NHUC. suggest that TERT-NHUC are not suitable for long-term tissue Comparison of the differentiation response in TERT-NHUC replacement strategies in patients. Our data provides support for soon after transduction to those that had undergone 100 PD the use of telomerase inhibitors in UC, because in addition to its found that low passage TERT-NHUC had a response similar to known actions at telomeres, it could be predicted that multiple that of mortal NHUC. Similarly, the increased saturation molecules would be targeted by a single intervention, leading to density observed in TERT-NHUC at 100 PD was not shown an immediate therapeutic effect. Further investigations into the soon after transduction with hTERT. Thus, attenuation of non–telomere effects of telomerase may provide valuable differentiation and cell-cell contact inhibition are not acute insights into processes relevant both during normal develop- effects of telomerase expression but occur after telomerase- ment and in cancer pathogenesis. mediated immortalization. The attenuation of these responses in TERT-NHUC at 100 PD suggests that caution is required if Materials and Methods using these cells as a platform to study gene function in normal Cell Lines urothelial cells. However, these cells may be a good model in TERT-NHUC A, TERT-NHUC B, and TERT-NHUC N and which to study gene function in the context of premalignancy. isogenic NHUC were cultured as described (2). Analyses of In summary, the identification of genes that were consis- TERT-NHUC were done on cells soon after transduction with tently altered after the expression of telomerase has led to hTERT (recovered from frozen samples within four passages of robust filtering of gene lists and the identification of potential selection and cultured for triplicate RNA extractions; PD level telomerase signature genes. We have confirmed that the was <18 PD) on cells that had undergone f100 PD and at a telomere maintenance effects of telomerase are required for final time point when cells had undergone at least an additional immortalization of NHUC. However, additional non–telomere 150 PD. Saturation density was assessed by plating 2 Â 104 effects include profound and consistent alterations in gene cells in duplicate 35-mm diameter dishes and counting after 2, expression which might be predicted to contribute to tumori- 5, 7, 9, 12, and 14 days. For conditioned medium experiments, genesis. Thus, further analysis of telomerase signature genes in medium was added to 100 PD cells at 50% confluence. After

FIGURE6. Expression of candidate telomerase signature genes was altered in UC cell lines and tumors. Log10 RQ relative to pooled NHUC cDNA and normalized to SDHA. The reference sample (pooled NHUC) therefore has a RQ value of = 1 and log10 RQ value of 0. Samples with a log10 RQ value < 0 have down-regulation of expression and those with log10 RQ > 0, overexpression relative to pooled NHUC. Where transcript was undetectable after 40 cycles of PCR, log10 RQ is shown as À5. Bladder cell lines from left to right are; CAL29, 253J, DSH1, HT1376, TCCSUP, J82, KU1919, RT112, JO’N, LUCC1, 97-7, 5637, 96-1, 97-6, SCaBER, SVHUC, 94-10, RT4, T24, UMUC3, JMSU, 97-1, SD, SW1170, 92-1, BFTC905, VMCUB2, and HCV29. SVHUC was derived from urothelium transformed in vitro, HCV29 was established from nonmalignant ureteric urothelium of a patient with bladder cancer and the remainder are UC cell lines. A. NDN, B, NME5, and C, ADFP. Expression of candidate telomerase -regulated genes was altered in primary UC. D, NDN; and E, NME5. Tumors are ranked in order of expression of each gene.

48 h, medium was harvested, centrifuged at 100 Â g for Hs00417200 (SDHA) and normalized to pooled NHUC cDNA. 10 min, filtered, and stored at À20jC. NHUC were seeded at TERT-NHUC RNA was similar to that used for array analyses. 6 Â 104 cells per 35 mm well in triplicate, fed every 3 days with Expression of NDN and NME5 was also examined in UC cell the conditioned medium and counted weekly. lines and tumors. cDNA was prepared from these cell lines as previously described (67). Microarray Processing and Data Analysis Affymetrix HG_U133 Plus 2.0 oligonucleotide arrays were Induction of Urothelial Differentiation hybridized at the Patterson Institute for Cancer Research, Cells were plated at 2 Â 105 cells/mL and when f70% Manchester, United Kingdom. Further information including confluent, cells were treated for 24 h with 1 Amol/L of RNA extraction methods is available at their web site.3 Two Troglitazone plus 1 Amol/L of the epidermal growth factor micrograms of total RNA was used to prepare biotinylated target receptor inhibitor, PD153035 (Merck; ref. 20), then maintained RNA, according to the Affymetrix One Cycle Target Preparation for 5 days in 1 Amol/L of PD153035. RNA was extracted using Protocol, driven by T7-linked oligo(dT) primers. RNA was GenElute mammalian total RNA mini-prep kit (Sigma) and extracted from TERT-NHUC, soon after transduction with cDNA was transcribed from 1 Ag of RNA using SuperScript hTERT (<18 PD), after 100 PD, and after f250 PD. Microarray First-Strand Synthesis system (Invitrogen). UPK2 expression data were analyzed using the Bioconductor (63) packages. cel was quantified using TaqMan QRT-PCR (assay Hs00171854). files were robust means analysis preprocessed (64) using the Limma package (65). A linear model was constructed to fit, Quantification of Telomerase Activity within each donor, a baseline NHUC level and offsets for Telomerase activity was measured using the TRAPeze-RT telomerase expression. The Limma package allows these kit (Chemicon) and Titanium Taq (Chemicon) according to the variables to then be tested, using an empirical Bayes method manufacturer’s instructions. Triplicate reactions containing that adjusts the per-gene replicate variance towards the global 1,000 cells per reaction were carried out on an ABI 7500. average variance across genes, to lessen the effect of variance The assay was repeated twice and an average value was underestimation. Differentially expressed genes were selected calculated. on the basis of the false discovery rate being controlled to 0.0001; this test was applied to discover separate gene lists for TERT-HA NHUC each comparison of TERT-NHUC to the baseline NHUC, within Retroviruses were produced using pBabe-puro vectors and cells from each donor. Gene lists were compared to identify ecotrophic packaging cells. NHUC expressing the ecotrophic genes consistently altered in cells from at least two of the three retroviral receptor (2) were transduced to express hTERT-HA, donors. Data were analyzed through the use of Ingenuity 4 wild-type hTERT, or empty vector (Addgene plasmids 1772, Pathway Analysis (Ingenuity Systems). Functional analysis 1771, and 1764).5 The expression of hTERT-HA leads to identified the biological functions and diseases that were most telomerase activity but, due to an HA tag on its COOH significant to the data set (list of genes altered in at least two of terminus, lacks the ability to act in telomere elongation or three donors). Genes from the data set that met the criteria of maintenance (15). >2-fold change and were associated with biological functions and/or diseases in the Ingenuity Pathways Knowledge Base were considered for the analysis. Fischer’s exact test was used to Disclosure of Potential Conflicts of Interest No potential conflicts of interest were disclosed. calculate a P value determining the probability that each biological function or disease assigned to the data set is due to Acknowledgments chance alone. For gene ontology analysis, Affymetrix probes We thank Stuart Pepper and the Cancer Research UK Affymetrix Facility team for were mapped to a unique set of Entrez IDs. GOstats (66) was their assistance with array experiments and Bob Weinberg’s laboratory for then used to test for overrepresentation at all the ‘‘biological plasmids deposited with Addgene. process’’ gene ontology nodes, using a hypergeometric test. References QRT-PCR 1. Muller M. Telomerase: its clinical relevance in the diagnosis of bladder cancer. Oncogene 2002;21:650 – 5. To validate microarray data, the expression of selected genes 2. Chapman EJ, Hurst CD, Pitt E, Chambers P, Aveyard JS, Knowles MA. was confirmed by QRT-PCR. RNA from triplicate repeats for Expression of hTERT immortalises normal human urothelial cells without each condition on the array was pooled. One microgram of total inactivation of the p16/Rbpathway. Oncogene 2006;25:5037 – 45. RNA was reverse-transcribed using Advantage RT-for-PCR kit 3. Smith LL, Coller HA, Roberts JM. Telomerase modulates expression of growth-controlling genes and enhances cell proliferation. Nat Cell Biol 2003;5: (Clontech). QRT-PCR was carried out using an ABI 7500 Real- 474 – 9. time PCR System and TaqMan Gene Expression Assays 4. Stewart SA, Hahn WC, O’Connor BF, et al. Telomerase contributes to (Applied Biosystems); Hs00267349 (NDN), Hs00177499 tumorigenesis by a telomere length-independent mechanism. Proc Natl Acad Sci (NME5), Hs00605340 (ADFP), Hs00180411 (BMI1), and U S A 2002;99:12606 – 11. Hs01009006 (SIRT1). Expression was quantified relative to 5. Zimmermann S, Martens UM. Telomeres and telomerase as targets for cancer therapy. Cell Mol Life Sci 2007;64:906 – 21. 6. Kraemer K, Fuessel S, Schmidt U, et al. Antisense-mediated hTERT inhibition specifically reduces the growth of human bladder cancer cells. Clin Cancer Res 3 http://bioinformatics.picr.man.ac-uk/mbcf/protocols.jsp 2003;9:3794 – 800. 4 http://www.ingenuity.com/ 7. Sharma HW, Sokoloski JA, Perez JR, et al. Differentiation of immortal cells 5 http://www.addgene.org/pgvec1 inhibits telomerase activity. Proc Natl Acad Sci U S A 1995;92:12343 – 6.

8. Bagheri S, Nosrati M, Li S, et al. Genes and pathways downstream of 32. Hashizume K, Nakajo T, Kawarasaki H, et al. Prader-Willi syndrome with telomerase in melanoma metastasis. Proc Natl Acad Sci U S A 2006;103: del(15)(q11,q13) associated with hepatoblastoma. Acta Paediatr Jpn 1991;33: 11306 – 11. 718 – 22. 9. Alge CS, Hauck SM, Priglinger SG, Kampik A, Ueffing M. Differential 33. Jaffray B, Moore L, Dickson AP. Prader-Willi syndrome and intratubular protein profiling of primary versus immortalized human RPE cells identifies germ cell neoplasia. Med Pediatr Oncol 1999;32:73 – 4. expression patterns associated with cytoskeletal remodeling and cell survival. 34. Taniura H, Taniguchi N, Hara M, Yoshikawa K. Necdin, a postmitotic J Proteome Res 2006;5:862 – 78. neuron-specific growth suppressor, interacts with viral transforming proteins and 10. Lindvall C, Hou M, Komurasaki T, et al. Molecular characterization of cellular transcription factor E2F1. J Biol Chem 1998;273:720 – 8. human telomerase reverse transcriptase-immortalized human fibroblasts by gene 35. Nakada Y, Taniura H, Uetsuki T, Yoshikawa K. Characterization and expression profiling: activation of the epiregulin gene. Cancer Res 2003;63: chromosomal mapping of a human Necdin pseudogene. Gene 2000;245:185 – 91. 1743 – 7. 36. Brunelli S, Tagliafico E, De Angelis FG, et al. Msx2 and necdin combined 11. Ramirez RD, Sheridan S, Girard L, et al. Immortalization of human activities are required for smooth muscle differentiation in mesoangioblast stem bronchial epithelial cells in the absence of viral oncoproteins. Cancer Res 2004; cells. Circ Res 2004;94:1571 – 8. 64:9027 – 34. 37. Tseng YH, Butte AJ, Kokkotou E, et al. Prediction of preadipocyte 12. Farwell DG, Shera KA, Koop JI, et al. Genetic and epigenetic changes in differentiation by gene expression reveals role of insulin receptor substrates and human epithelial cells immortalized by telomerase. Am J Pathol 2000;156: necdin. Nat Cell Biol 2005;7:601 – 11. 1537 – 47. 38. Ohman Forslund K, Nordqvist K. The melanoma antigen genes—any clues 13. Belair CD, Yeager TR, Lopez PM, Reznikoff CA. Telomerase activity: a to their functions in normal tissues? Exp Cell Res 2001;265:185 – 94. biomarker of cell proliferation, not malignant transformation. Proc Natl Acad Sci U S A 1997;94:13677 – 82. 39. Sigala S, Faraoni I, Botticini D, et al. Suppression of telomerase, reexpression of KAI1, and abrogation of tumorigenicity by nerve growth factor 14. Masutomi K, Yu EY, Khurts S, et al. Telomerase maintains telomere structure in prostate cancer cell lines. Clin Cancer Res 1999;5:1211 – 8. in normal human cells. Cell 2003;114:241 – 53. 40. Tabassum A, Khwaja F, Djakiew D. The p75(NTR) tumor suppressor 15. Counter CM, Hahn WC, Wei W, et al. Dissociation among in vitro induces caspase-mediated apoptosis in bladder tumor cells. Int J Cancer 2003; telomerase activity, telomere maintenance, and cellular immortalization. Proc Natl 105:47 – 52. Acad Sci U S A 1998;95:14723 – 8. 41. Taniura H, Matsumoto K, Yoshikawa K. Physical and functional 16. Bracken AP, Dietrich N, Pasini D, Hansen KH, Helin K. Genome-wide interactions of neuronal growth suppressor necdin with p53. J Biol Chem mapping of Polycombtarget genes unravels their roles in cell fate transitions. 1999;274:16242 – 8. Genes Dev 2006;20:1123 – 36. 42. Taniura H, Kobayashi M, Yoshikawa K. Functional domains of necdin for 17. Maruyama K, Usami M, Aizawa T, Yoshikawa K. A novel brain-specific protein-protein interaction, nuclear matrix targeting, and cell growth suppression. mRNA encoding nuclear protein (necdin) expressed in neurally differentiated J Cell Biochem 2005;94:804 – 15. embryonal carcinoma cells. Biochem Biophys Res Commun 1991;178:291 – 6. 43. Matsumoto K, Taniura H, Uetsuki T, Yoshikawa K. Necdin acts as a 18. Kuwako K, Taniura H, Yoshikawa K. Necdin-related MAGE proteins transcriptional repressor that interacts with multiple guanosine clusters. Gene differentially interact with the E2F1 transcription factor and the p75 neurotrophin 2001;272:173 – 9. receptor. J Biol Chem 2004;279:1703 – 12. 44. Gil J, Bernard D, Peters G. Role of polycombgroup proteins in stem cell self- 19. Nowak K, Kerl K, Fehr D, et al. BMI1 is a target gene of E2F-1 and is renewal and cancer. DNA Cell Biol 2005;24:117 – 25. strongly expressed in primary neuroblastomas. Nucleic Acids Res 2006;34: 1745 – 54. 45. Bracken AP, Kleine-Kohlbrecher D, Dietrich N, et al. The Polycomb group proteins bind throughout the INK4A-ARF locus and are disassociated in 20. Varley CL, Stahlschmidt J, Lee WC, et al. Role of PPARg and EGFR senescent cells. Genes Dev 2007;21:525 – 30. signalling in the urothelial terminal differentiation programme. J Cell Sci 2004; 117:2029 – 36. 46. van Leenders GJ, Dukers D, Hessels D, et al. Polycomb-group oncogenes EZH2, BMI1, and RING1 are overexpressed in prostate cancer with adverse 21. Harnden P, Allam A, Joyce AD, Patel A, Selby P, Southgate J. Cytokeratin pathologic and clinical features. Eur Urol 2007;52:455 – 63. 20 expression by non-invasive transitional cell carcinomas: potential for distinguishing recurrent from non-recurrent disease. Histopathology 1995;27: 47. Tateishi K, Ohta M, Kanai F, et al. Dysregulated expression of stem cell 169 – 74. factor Bmi1 in precancerous lesions of the gastrointestinal tract. Clin Cancer Res 2006;12:6960 – 6. 22. Varley CL, Stahlschmidt J, Smith B, Stower M, Southgate J. Activation of peroxisome proliferator-activated receptor-g reverses squamous metaplasia and 48. Sparmann A, van Lohuizen M. Polycombsilencers control cell fate, induces transitional differentiation in normal human urothelial cells. Am J Pathol development and cancer. Nat Rev Cancer 2006;6:846 – 56. 2004;164:1789 – 98. 49. Glinsky GV, Berezovska O, Glinskii AB. Microarray analysis identifies a 23. Southgate J, Hutton KA, Thomas DF, Trejdosiewicz LK. Normal human death-from-cancer signature predicting therapy failure in patients with multiple urothelial cells in vitro: proliferation and induction of stratification. LabInvest types of cancer. J Clin Invest 2005;115:1503 – 21. 1994;71:583 – 94. 50. Nosrati M, Li S, Bagheri S, et al. Antitumor activity of systemically 24. Choi J, Southworth LK, Sarin KY, et al. TERT promotes epithelial delivered ribozymes targeting murine telomerase RNA. Clin Cancer Res 2004; 10:4983 – 90. proliferation through transcriptional control of a Myc- and Wnt-related developmental program. PLoS Genet 2008;4:e10. 51. Liu L, Berletch JB, Green JG, Pate MS, Andrews LG, Tollefsbol TO. Telomerase inhibition by retinoids precedes cytodifferentiation of leukemia 25. Ringrose L. Polycombcomes of age: genome-wide profiling of target sites. cells and may contribute to terminal differentiation. Mol Cancer Ther 2004;3: Curr Opin Cell Biol 2007;19:290 – 7. 1003 – 9. 26. Kuzmichev A, Margueron R, Vaquero A, et al. Composition and histone 52. Richardson RM, Nguyen B, Holt SE, Broaddus WC, Fillmore HL. Ectopic substrates of polycomb repressive group complexes change during cellular telomerase expression inhibits neuronal differentiation of NT2 neural progenitor differentiation. Proc Natl Acad Sci U S A 2005;102:1859 – 64. cells. Neurosci Lett 2007;421:168 – 72. 27. Furuyama T, Banerjee R, Breen TR, Harte PJ. SIR2 is required for polycomb 53. Okegawa T, Pong RC, Li Y, Bergelson JM, Sagalowsky AI, Hsieh JT. The silencing and is associated with an E(Z) histone methyltransferase complex. Curr mechanism of the growth-inhibitory effect of coxsackie and adenovirus receptor Biol 2004;14:1812 – 21. (CAR) on human bladder cancer: a functional analysis of car protein structure. 28. Pruitt K, Zinn RL, Ohm JE, et al. Inhibition of SIRT1 reactivates silenced Cancer Res 2001;61:6592 – 600. cancer genes without loss of promoter DNA hypermethylation. PLoS Genet 2006; 54. Giuliani L, Ciotti M, Stoppacciaro A, et al. UDP-glucuronosyltransferases 2:e40. 1A expression in human urinary bladder and colon cancer by immunohisto- 29. Vaquero A, Scher M, Erdjument-Bromage H, Tempst P, Serrano L, Reinberg chemistry. Oncol Rep 2005;13:185 – 91. D. SIRT1 regulates the histone methyl-transferase SUV39H1 during heterochro- 55. Bahrenberg G, Brauers A, Joost HG, Jakse G. Reduced expression of PSCA, matin formation. Nature 2007;450:440 – 4. a member of the LY-6 family of cell surface antigens, in bladder, esophagus, and 30. Davies HD, Leusink GL, McConnell A, et al. Myeloid leukemia in Prader- stomach tumors. Biochem Biophys Res Commun 2000;275:783 – 8. Willi syndrome. J Pediatr 2003;142:174 – 8. 56. Tang S, Bhatia B, Maldonado CJ, et al. Evidence that arachidonate 31. Coppes MJ, Sohl H, Teshima IE, Mutirangura A, Ledbetter DH, Weksberg R. 15-lipoxygenase 2 is a negative cell cycle regulator in normal prostate epithelial Wilms tumor in a patient with Prader-Willi syndrome. J Pediatr 1993;122:730 – 3. cells. J Biol Chem 2002;277:16189 – 201.

57. Takazaki R, Nishimura I, Yoshikawa K. Necdin is required for terminal dimensional model of differentiation of immortalized human bronchial epithelial differentiation and survival of primary dorsal root ganglion neurons. Exp Cell Res cells. Differentiation 2006;74:141 – 8. 2002;277:220 – 32. 63. Gentleman RC, Carey VJ, Bates DM, et al. Bioconductor: open software 58. Chow NH, Liu HS, Chan SH. The role of nm23-1 in the progression of development for computational biology and bioinformatics. Genome Biol 2004; transitional cell bladder cancer. Clin Cancer Res 2000;6:3595 – 9. 5:R80. 59. Munier A, Serres C, Kann ML, et al. Nm23/NDP kinases in human male 64. Irizarry RA, Hobbs B, Collin F, et al. Exploration, normalization, and germ cells: role in spermiogenesis and sperm motility? Exp Cell Res 2003;289: summaries of high density oligonucleotide array probe level data. Biostatistics 295 – 306. 2003;4:249 – 64. 60. Al-Tubuly AA, Spijker R, Pignatelli M, Kirkland SC, Ritter MA. Inhibition 65. Smyth GK. Limma: linear models for microarray data. New York: Springer; of growth and enhancement of differentiation of colorectal carcinoma cell lines by 2005. MAbMR6 and IL-4. Int J Cancer 1997;71:605 – 11. 66. Gentleman R. Using GO for statistical analyses. Compstat 2004 Proceedings 61. Yao M, Huang Y, Shioi K, et al. Expression of adipose differentiation-related in Computational Statistics; 2004. protein: a predictor of cancer-specific survival in clear cell renal carcinoma. Clin 67. Tomlinson DC, Hurst CD, Knowles MA. Knockdown by shRNA identifies Cancer Res 2007;13:152 – 60. S249C mutant FGFR3 as a potential therapeutic target in bladder cancer. 62. Vaughan MB, Ramirez RD, Wright WE, Minna JD, Shay JW. A three- Oncogene 2007;26:5889 – 99.

Mol Cancer Res 2008;6(7). July 2008 Downloaded from mcr.aacrjournals.org on October 1, 2021. © 2008 American Association for Cancer Research. Genes Involved in Differentiation, Stem Cell Renewal, and Tumorigenesis Are Modulated in Telomerase-Immortalized Human Urothelial Cells

Emma J. Chapman, Gavin Kelly and Margaret A. Knowles

Mol Cancer Res 2008;6:1154-1168.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/6/7/1154

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2008/07/25/6.7.1154.DC1

Cited articles This article cites 64 articles, 24 of which you can access for free at: http://mcr.aacrjournals.org/content/6/7/1154.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/6/7/1154.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/6/7/1154. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.